This is basically a hand-holding session for
a friend with very little knowledge of unix
or DSP at the time. I've left it here in
the off chance that it may be useful to
someone else in that situation. If you're
hip, you may find this file too cloying
for your taste.

---
I've uploaded one of my synthesizers. It's basically a sonic
lava lamp. Following are instructions for use. All commands
that you should type in are preceeded by the "wendy>" prompt.


In order to run the program, type the following:

wendy> jones -d10 | float2fix | tosnd -h /scratch/erik/ambient1.snd &

This may take a minute or so. The ampersand (&) puts the
job in the background, and jones will print a message
when it has concluded. Then just click on the soundfile
to hear it. (Note - do not try to play the sound before
computation has concluded, as this may well crash the
NeXT computer.)

A brief explanation of what you just did:
you ran jones, which creates sound data, piped the
output to fix2float, which changes the format of the
samples into the format for playback, and finally
piped this through tosnd, which puts a NeXT header
on the soundfile. Let me call your attention to
the constructions -d10 and -h. These are what are
known as flags, and are used to send information to
programs on the command line. Notice that there are no
spaces allowed in these flags. the "-" sign indicates a flag.
The letter which follows is the name of the flag.
the value following the flag name is the data to be
passed to the program. Notice that the d flag is followed
by 10 (indicating a duration of 10 seconds) and the
h flag has no value. (h indicates 44100 sampling rate.)
For more info on tosnd or fix2float, just type

wendy> tosnd
or
wendy> fix2float

with no arguments and the program will print a list
of options. My program also does this. If you type:

wendy> jones

you will see the following:

jones:  baroque random spectral synthesizer
jones   [flags]  > floatsams
        R:      sampling rate [44100]
        N:      fft length [1024]
        M:      window size in samples [4096]
        I:      interpolation factor in samples [512]
        l:      minimum spectral duration [.1]
        u:      maximum spectral duration [.8]
        d:      duration [undefined]
        o:      odds against amplitude appearing [.95]
        f:      scaling for fine components[.1]
        m:      optional function for rescaling spectrum
        x:      bin frequency deviation percentage [0.01]
        s:      optional seed

Notice the line
        d:      duration [undefined]
	
which indicates that duration is passed through
the d flag, and that it does not have a default
value. All the other flags have default values so
you don't have to understand them all immediately
before using the program. But if you like the program,
I can explain them all eventually. One set which is
easy to play with is the u and l flags, which specify
the minimum and maximum duration of segments, thus 
controlling the speed of sound evolution. To make a sound 
which evolves more slowly you might type:

jones -d10 -l.5 -u3.0 | float2fix | tosnd -h /scratch/erik/ambient2.snd &

That's all for now. Let me know how it works out,
and have fun.
